Differentiation And Research On The Measurement Of The Features Of Flat-granular Object

The image recognizing and processing technology was first put into practices in western developed countries since 1970's, which has great influence on the efficiency of labor. It has been applied into more fields since then. The sorting of flat-granular objects is related to many fields, e.g. the image recognition, Optical imaging, industry control and Mechanical Drive. The flat-granular objects are tiny, irregular and difficult with little information. The researchs on the flat-granular objects are very limited.This paper focuses on the feature extraction of the flat-granular objects. It aims to improve the efficiency and accuracy of the sorting and develop an original system of laboratory sorting, which is finally applied into other fields. The objects are:1. Balance between the efficiency and accuracy of the sorting. Research and analyze the algorithm of image prehandling. Find the proper method and meet the demand of real-time.2. Put forward new design by extracting characteristic values, such as pattern connectivity and black region coverage and gray histogram distribution difference. And the new design performs perfect in practice application.3. Put forward a new algorithm to measure the dimension of the flat-granular objects, which solves the disadvantage of the elder algorithms, which are ineffective and have more errors. The new algorithm can be applied into other relevant measurement.4. Introduce a way to measure the flat-granular objects, to decide wheather its shape is regular or not.5. Design a way to sort the flat-granular objects, which is practical, cheap and easy to meet the demands of sorting as well as efficient.